DC2

park_movement_Sun <- read_csv("Movement Data/park-movement-Sun.csv", 
    col_types = cols(Timestamp = col_datetime(format = "%Y-%m-%d %H:%M:%S\t")))

## Warning: 5 parsing failures.
##      row       col                      expected    actual                                  file
##  4332993 NA        5 columns                     1 columns 'Movement Data/park-movement-Sun.csv'
## 10932426 Timestamp date like %Y-%m-%d %H:%M:%S    Timestamp 'Movement Data/park-movement-Sun.csv'
## 10932426 id        a double                      id        'Movement Data/park-movement-Sun.csv'
## 10932426 X         a double                      X         'Movement Data/park-movement-Sun.csv'
## 10932426 Y         a double                      Y         'Movement Data/park-movement-Sun.csv'

# 2014-6-08 08:00:11    
park_movement_Sat <- read_csv("Movement Data/park-movement-Sat.csv",
    col_types = cols(Timestamp = col_datetime(format = "%Y-%m-%D %H:%M:%S\t")))
# 2014-6-07 08:00:08    
park_movement_Fri <- read_csv("Movement Data/park-movement-Fri.csv",
    col_types = cols(Timestamp = col_datetime(format = "%Y-%m-%D %H:%M:%S\t")))
# 2014-6-06 08:00:19    

Time

park_movement_Fri_n <- park_movement_Fri %>% group_by(Timestamp) %>%
  mutate(n = 1, count = sum(n)) %>%
  select(id,Timestamp,count,type) %>% distinct(Timestamp, .keep_all=T)
park_movement_Sat_n <- park_movement_Sat %>% group_by(Timestamp) %>%
  mutate(n = 1, count = sum(n))%>% 
  select(Timestamp,count,type) %>% distinct(Timestamp, .keep_all=T) 
park_movement_Sun_n <- park_movement_Sun %>%  group_by(Timestamp) %>%
  mutate(n = 1, count = sum(n))%>% select(Timestamp,count,type) %>% distinct(Timestamp, .keep_all=T) 

fri_n <- ggplot(data = park_movement_Fri_n, aes(x=Timestamp, y=count)) + geom_line()+
  scale_x_datetime(date_breaks = "1 hour") + labs(x="Friday",y=NULL) + theme_bw() +
  theme(axis.text.x = element_blank())
sat_n <- ggplot() + geom_line(data = park_movement_Sat_n, aes(x=Timestamp, y=count))+
  scale_x_datetime(date_breaks = "1 hour") + labs(x="Saturday",y=NULL) + theme_bw()+
  theme(axis.text.x = element_blank())
sun_v <- ggplot() + geom_line(data = park_movement_Sun_n, aes(x=Timestamp, y=count)) +
  scale_x_datetime(date_breaks = "1 hour",date_labels = "%H:%M")+ labs(x="Sunday",y=NULL) + theme_bw()

ggpubr::ggarrange(fri_n,sat_n,sun_v, nrow = 3)

## Warning: Removed 1 row(s) containing missing values (geom_path).

Time check in

park_movement_Fri_n_check <- park_movement_Fri %>% filter(type == "check-in") %>% 
  mutate(hour = substr(Timestamp, 12,13)) %>% group_by(hour) %>% mutate(n = 1, count = sum(n))%>%
  distinct(hour, .keep_all=T)%>%
  select(Timestamp,count,type)

## Adding missing grouping variables: `hour`

park_movement_Sat_n_check <- park_movement_Sat %>% filter(type == "check-in") %>%
  mutate(hour = substr(Timestamp, 12,13)) %>% group_by(hour) %>% mutate(n = 1, count = sum(n))%>%
  distinct(hour, .keep_all=T)%>%
  select(Timestamp,count,type)

## Adding missing grouping variables: `hour`

park_movement_Sun_n_check <- park_movement_Sun %>% filter(type == "check-in") %>%
  mutate(hour = substr(Timestamp, 12,13)) %>% group_by(hour) %>% mutate(n = 1, count = sum(n))%>%
  distinct(hour, .keep_all=T)%>%
  select(Timestamp,count,type)

## Adding missing grouping variables: `hour`

fri_n_c <- ggplot(data = park_movement_Fri_n_check, aes(x=as.integer(hour), y=count)) + geom_line()+
  theme_bw() +
  scale_x_discrete(name = "Friday",limits=c(8:23))+
  theme(axis.text.x = element_blank(),
        axis.ticks = element_blank()) + labs(y=NULL)

## Warning: Continuous limits supplied to discrete scale.
## Did you mean `limits = factor(...)` or `scale_*_continuous()`?

sat_n_c <- ggplot() + geom_line(data = park_movement_Sat_n_check, aes(x=as.integer(hour), y=count))+
  theme_bw() +
  scale_x_discrete(name = "Saturday",limits=c(8:23)) +
  theme(axis.text.x = element_blank(),
        axis.ticks = element_blank())+ labs(y=NULL)

## Warning: Continuous limits supplied to discrete scale.
## Did you mean `limits = factor(...)` or `scale_*_continuous()`?

sun_v_c <- ggplot() + geom_line(data = park_movement_Sun_n_check, aes(x=as.integer(hour), y=count))+
  theme_bw() +
  scale_x_discrete(name = "Sunday",limits=c(8:23)) + theme(axis.title.y = element_blank()) 

## Warning: Continuous limits supplied to discrete scale.
## Did you mean `limits = factor(...)` or `scale_*_continuous()`?

ggpubr::ggarrange(fri_n_c,sat_n_c,sun_v_c, nrow = 3)

park_movement_Fri_sl <- park_movement_Fri%>% group_by(Timestamp) %>% mutate(n=1, count=sum(n)) %>%
  distinct(Timestamp,.keep_all=T) %>%
  group_by(id) %>% mutate(stay_length = sum(n)) %>%
  arrange(desc(stay_length)) %>%  select(Timestamp, id, stay_length, X,Y)

park_movement_Sat_sl <- park_movement_Sat%>% group_by(Timestamp) %>% mutate(n=1, count=sum(n)) %>%
  distinct(Timestamp,.keep_all=T) %>%
  group_by(id) %>% mutate(stay_length = sum(n)) %>%
  arrange(desc(stay_length))  %>%  select(Timestamp, id, stay_length, X,Y)

park_movement_Sun_sl <- park_movement_Sun%>% group_by(Timestamp) %>% mutate(n=1, count=sum(n)) %>%
  distinct(Timestamp,.keep_all=T) %>%
  group_by(id) %>% mutate(stay_length = sum(n)) %>%
  arrange(desc(stay_length))  %>% select(Timestamp, id, stay_length, X,Y)

park_movement_Fri_id <- park_movement_Fri_sl %>% distinct(id,.keep_all=T) %>% 
  select(id, stay_length)

park_movement_Sat_id <- park_movement_Sat_sl %>% distinct(id,.keep_all=T) %>% 
  select(id, stay_length)

park_movement_Sun_id <- park_movement_Sun_sl %>% distinct(id,.keep_all=T) %>% 
  select(id, stay_length)

sus_id_sl <- park_movement_Fri_id%>%
  full_join(park_movement_Sat_id) %>%
  full_join(park_movement_Sun_id) %>%
  mutate(n=1) %>% group_by(id) %>% 
  mutate(many_visits=sum(n), total_stay_length = sum(stay_length)) %>%
  arrange(desc(total_stay_length)) %>% filter(many_visits>=2) %>% distinct(id, .keep_all =T) %>%
  select(id, many_visits, total_stay_length)

## Joining, by = c("id", "stay_length")
## Joining, by = c("id", "stay_length")

# write_csv(movement_id_sl, "sus-id.csv")

# ggplot(data = park_movement_Fri_id, aes(x=Timestamp, y=count)) + geom_line()

id_list <- as.vector(unlist(sus_id_sl[c(1:50),1]))

Movement Map for IDs

Friday

sus_movement_fri <- park_movement_Fri%>%
  filter(id %in% id_list) 

g1 <- ggplot(data=sus_movement_fri, aes(x=X, y=Y, color=as.character(id))) +geom_point(alpha=0.3)
g1

# ggplotly(g1)

Saturday

sus_movement_sat <- park_movement_Sat %>%
  filter(id %in% id_list) #, Timestamp >= "") 

g3 <- ggplot(data=sus_movement_sat, aes(x=X, y=Y, color=as.character(id))) +geom_point(alpha=0.3)
g3

Sunday

sus_movement_sun <- park_movement_Sun %>%
  filter(id %in% id_list)

g5 <- ggplot(data=sus_movement_sun, aes(x=X, y=Y, color=as.character(id))) +geom_point(alpha=0.3)
g5

Further narrow by location

id_close_fri <- park_movement_Fri%>%
  filter(X==32&Y==33& type=="check-in") %>% distinct(id) %>% select(id) %>% mutate(day = "Fri")

id_close_sat <- park_movement_Sat%>%
  filter(X==32&Y==33& type=="check-in") %>% distinct(id) %>% select(id) %>% mutate(day = "Sat")

id_close_sun <- park_movement_Sun%>%
  filter(X==32&Y==33& type=="check-in") %>% distinct(id) %>% select(id) %>% mutate(day = "Sun")

id_close <- id_close_fri%>% 
  full_join(id_close_sat) %>% full_join(id_close_sun) %>% 
  mutate(n=1) %>% group_by(id) %>% mutate(times = sum(n)) %>% distinct()  %>%
  filter(times==3) %>% select (id, times)

## Joining, by = c("id", "day")
## Joining, by = c("id", "day")

write_csv(id_close, "id_close")  #5,439 rows --> 2,457 r (both check in and movement) --> 480 (checkin only)

Location + stay length

id_sus <- id_close %>%
  full_join(sus_id_sl) %>% na.omit() %>% distinct(id, .keep_all=T) #153 rows --> 15 checkin

## Joining, by = "id"

id_sus_15 <- as.vector(unlist(id_sus[,1]))

Communication Data

comm_data_Fri <- read_csv("Communication Data/comm-data-Fri.csv", 
                          col_types = cols(Timestamp = col_datetime(format = "%Y-%m-%d %H:%M:%S\t"),
                                           from = col_character()))
comm_data_Sat <- read_csv("Communication Data/comm-data-Sat.csv", 
                          col_types = cols(Timestamp = col_datetime(format = "%Y-%m-%d %H:%M:%S\t"),
                                           from = col_character()))
comm_data_Sun <- read_csv("Communication Data/comm-data-Sun.csv", 
                          col_types = cols(Timestamp = col_datetime(format = "%Y-%m-%d %H:%M:%S\t"),
                                           from = col_character()))

Filter by movement + staylength id list

id_sus <- as.vector(unlist(sus_id_sl[,1]))
  
comm_data_Fri_f <- comm_data_Fri%>%
  filter(from %in% id_sus|to %in% id_sus) %>%
  group_by(Timestamp)%>%
  mutate(m=1, n = sum(m))
comm_data_Sat_f <- comm_data_Sat%>%
  filter(from %in% id_sus|to %in% id_sus)%>%
  group_by(Timestamp)%>%
  mutate(m=1, n = sum(m))
comm_data_Sun_f <- comm_data_Sun%>%
  filter(from %in% id_sus|to %in% id_sus)%>%
  group_by(Timestamp)%>%
  mutate(m=1, n = sum(m))

## Friday
ggplot(data = comm_data_Fri_f, aes(x=Timestamp, y = n, color = from)) + geom_point(alpha=0.2) + 
  labs(x="Friday") + theme(legend.position="none")

## Satday
ggplot(data = comm_data_Sat_f, aes(x=Timestamp, y = n, color = from)) + geom_point(alpha=0.2) +
  labs(x="Saturday") + theme(legend.position="none")

## Sunday
ggplot(data = comm_data_Sun_f, aes(x=Timestamp, y = n, color = from)) + geom_point(alpha=0.2) + 
  labs(x="Sunday") + theme(legend.position="none")

Filter by IDs and External numbers

ext_id <- comm_data_Sun %>%
  filter(to == "external") %>% distinct(from)

Network graph

On 15 IDs idetified in terms of staylength and entry into the pavillion all 3 days

# id_sus <- id_sus %>%
#   full_join(ext_id)
# morning sunday: until 2014-06-08 13:16:27 

morning_sun_comm <- comm_data_Sun[c(1:527059),]
netw <- morning_sun_comm%>%
  filter(from%in%id_sus_15) %>% select(-location) # or to%in%id_sus_15|from%in%id_sus_15

#node list
sources <- netw %>% distinct(from) %>% rename(id = from) %>% mutate(from = "from")
destinations <- netw %>% distinct(to) %>% rename(id = to)
nodes <- full_join(sources, destinations, by = "id") %>% distinct(id,.keep_all=T) %>% 
  mutate(f = ifelse(from=="from", "from", "to")) %>% select(id,f)

#edge list
per_route <- netw %>%  
  group_by(from, to) %>%
  summarise(weight = n()) %>% 
  ungroup()

## `summarise()` has grouped output by 'from'. You can override using the `.groups` argument.

edges <- per_route %>% 
  left_join(nodes, by = c("from" = "id"))
edges <- edges %>% 
  left_join(nodes, by = c("to" = "id"))
edges <- select(edges, from, to, weight)

# A tbl_graph consists of two tibbles: an edges tibble and a nodes tibble
routes_tidy <- tbl_graph(nodes = nodes, edges = edges, directed = TRUE) 

nw <- ggraph(routes_tidy, layout = "auto") + 
  geom_edge_link(aes(color = weight, width = weight), alpha = 0.7) + 
  scale_edge_width(range = c(0.2, 2)) +
  scale_edge_colour_gradient(low = "#f09c9c",
                             high = "#d60d0d",
                             space = "Lab") +
  geom_node_point(aes(colour = ifelse(f=="from" ,"#0f0fa3","#635d5d")), show.legend = FALSE) + theme_graph()+
  geom_node_text(aes(label = ifelse(f=="from", id, NA)), repel = TRUE) 

## Using `stress` as default layout

nw

## Warning: Removed 382 rows containing missing values (geom_text_repel).

Misc

# until 2014-08-01 13:03:22 
f <- park_movement_Sun[c(1:4165607),] %>%
  filter(type=="check-in") %>% group_by(Timestamp) %>%
  mutate(n=1,count=sum(n)) %>% distinct(Timestamp,.keep_all=T)
ggplot(data =f, aes(x=Timestamp, y= count)) + geom_line()

s1128580 <- park_movement_Sun[c(1:4165607),]%>%
  filter(id=="1128580")
s1128580_n <- park_movement_Sun[c(4165607:10932426),]%>%
  filter(id=="1128580")
g3 <- ggplot(data=s1128580, aes(x=X, y=Y, color=Timestamp)) +
  geom_point(alpha=0.3) +labs (x = NULL, y =NULL, title = "ID 1128580's morning movement") + theme_classic() +
  theme(legend.position = "none") 
g4 <- ggplot(data=s1128580_n, aes(x=X, y=Y, color=Timestamp)) +
  geom_point(alpha=0.3) +labs (x = NULL, y =NULL,title = "ID 1128580's afternoon movement") + theme_classic()+
  theme(legend.position = "none")

ggpubr::ggarrange(g3,g4, ncol = 2, legend = "bottom")